Plant Phenomics最新文献

The development of unmanned aerial vehicle (UAV) remote sensing has been increasingly applied in forestry for high-throughput and rapid acquisition of tree phenomics traits for various research areas. However, the detection of individual trees and the extraction of their spectral data remain a challenge, often requiring manual annotation. Although several software-based solutions have been developed, they are far from being widely adopted. This paper presents ExtSpecR, an open-source tool for spectral extraction of a single tree in forestry with an easy-to-use interactive web application. ExtSpecR reduces the time required for single tree detection and annotation and simplifies the entire process of spectral and spatial feature extraction from UAV-based imagery. In addition, ExtSpecR provides several functionalities with interactive dashboards that allow users to maximize the quality of information extracted from UAV data. ExtSpecR can promote the practical use of UAV remote sensing data among forest ecology and tree breeding researchers and help them to further understand the relationships between tree growth and its physiological traits.

Rice (Oryza sativa) is an essential stable food for many rice consumption nations in the world and, thus, the importance to improve its yield production under global climate changes. To evaluate different rice varieties' yield performance, key yield-related traits such as panicle number per unit area (PNpM²) are key indicators, which have attracted much attention by many plant research groups. Nevertheless, it is still challenging to conduct large-scale screening of rice panicles to quantify the PNpM² trait due to complex field conditions, a large variation of rice cultivars, and their panicle morphological features. Here, we present Panicle-Cloud, an open and artificial intelligence (AI)-powered cloud computing platform that is capable of quantifying rice panicles from drone-collected imagery. To facilitate the development of AI-powered detection models, we first established an open diverse rice panicle detection dataset that was annotated by a group of rice specialists; then, we integrated several state-of-the-art deep learning models (including a preferred model called Panicle-AI) into the Panicle-Cloud platform, so that nonexpert users could select a pretrained model to detect rice panicles from their own aerial images. We trialed the AI models with images collected at different attitudes and growth stages, through which the right timing and preferred image resolutions for phenotyping rice panicles in the field were identified. Then, we applied the platform in a 2-season rice breeding trial to valid its biological relevance and classified yield production using the platform-derived PNpM² trait from hundreds of rice varieties. Through correlation analysis between computational analysis and manual scoring, we found that the platform could quantify the PNpM² trait reliably, based on which yield production was classified with high accuracy. Hence, we trust that our work demonstrates a valuable advance in phenotyping the PNpM² trait in rice, which provides a useful toolkit to enable rice breeders to screen and select desired rice varieties under field conditions.

The environmental conditions in customered speed breeding practice are, to some extent, empirical and, thus, can be further optimized. Crop and plant models have been developed as powerful tools in predicting growth and development under various environments for extensive crop species. To improve speed breeding, crop models can be used to predict the phenotypes resulted from genotype by environment by management at the population level, while plant models can be used to examine 3-dimensional plant architectural development by microenvironments at the organ level. By justifying the simulations via numerous virtual trials using models in testing genotype × environment × management, an optimized combination of environmental factors in achieving desired plant phenotypes can be quickly determined. Artificial intelligence in assisting for optimization is also discussed. We admit that the appropriate modifications on modeling algorithms or adding new modules may be necessary in optimizing speed breeding for specific uses. Overall, this review demonstrates that crop and plant models are promising tools in providing the optimized combinations of environment factors in advancing crop growth and development for speed breeding.

Stomata play an essential role in regulating water and carbon dioxide levels in plant leaves, which is important for photosynthesis. Previous deep learning-based plant stomata detection methods are based on horizontal detection. The detection anchor boxes of deep learning model are horizontal, while the angle of stomata is randomized, so it is not possible to calculate stomata traits directly from the detection anchor boxes. Additional processing of image (e.g., rotating image) is required before detecting stomata and calculating stomata traits. This paper proposes a novel approach, named DeepRSD (deep learning-based rotating stomata detection), for detecting rotating stomata and calculating stomata basic traits at the same time. Simultaneously, the stomata conductance loss function is introduced in the DeepRSD model training, which improves the efficiency of stomata detection and conductance calculation. The experimental results demonstrate that the DeepRSD model reaches 94.3% recognition accuracy for stomata of maize leaf. The proposed method can help researchers conduct large-scale studies on stomata morphology, structure, and stomata conductance models.

Abiotic stresses such as heat and frost limit plant growth and productivity. Image-based field phenotyping methods allow quantifying not only plant growth but also plant senescence. Winter crops show senescence caused by cold spells, visible as declines in leaf area. We accurately quantified such declines by monitoring changes in canopy cover based on time-resolved high-resolution imagery in the field. Thirty-six winter wheat genotypes were measured in multiple years. A concept termed "frost damage index" (FDI) was developed that, in analogy to growing degree days, summarizes frost events in a cumulative way. The measured sensitivity of genotypes to the FDI correlated with visual scorings commonly used in breeding to assess winter hardiness. The FDI concept could be adapted to other factors such as drought or heat stress. While commonly not considered in plant growth modeling, integrating such degradation processes may be key to improving the prediction of plant performance for future climate scenarios.

Accurate counting of maize tassels is essential for monitoring crop growth and estimating crop yield. Recently, deep-learning-based object detection methods have been used for this purpose, where plant counts are estimated from the number of bounding boxes detected. However, these methods suffer from 2 issues: (a) The scales of maize tassels vary because of image capture from varying distances and crop growth stage; and (b) tassel areas tend to be affected by occlusions or complex backgrounds, making the detection inefficient. In this paper, we propose a multiscale lite attention enhancement network (MLAENet) that uses only point-level annotations (i.e., objects labeled with points) to count maize tassels in the wild. Specifically, the proposed method includes a new multicolumn lite feature extraction module that generates a scale-dependent density map by exploiting multiple dilated convolutions with different rates, capturing rich contextual information at different scales more effectively. In addition, a multifeature enhancement module that integrates an attention strategy is proposed to enable the model to distinguish between tassel areas and their complex backgrounds. Finally, a new up-sampling module, UP-Block, is designed to improve the quality of the estimated density map by automatically suppressing the gridding effect during the up-sampling process. Extensive experiments on 2 publicly available tassel-counting datasets, maize tassels counting and maize tassels counting from unmanned aerial vehicle, demonstrate that the proposed MLAENet achieves marked advantages in counting accuracy and inference speed compared to state-of-the-art methods. The model is publicly available at https://github.com/ShiratsuyuShigure/MLAENet-pytorch/tree/main.

Plant phenomics aims to perform high-throughput, rapid, and accurate measurement of plant traits, facilitating the identification of desirable traits and optimal genotypes for crop breeding. Salvia miltiorrhiza (Danshen) roots possess remarkable therapeutic effect on cardiovascular diseases, with huge market demands. Although great advances have been made in metabolic studies of the bioactive metabolites, investigation for S. miltiorrhiza roots on other physiological aspects is poor. Here, we developed a framework that utilizes image feature extraction software for in-depth phenotyping of S. miltiorrhiza roots. By employing multiple software programs, S. miltiorrhiza roots were described from 3 aspects: agronomic traits, anatomy traits, and root system architecture. Through K-means clustering based on the diameter ranges of each root branch, all roots were categorized into 3 groups, with primary root-associated key traits. As a proof of concept, we examined the phenotypic components in a series of randomly collected S. miltiorrhiza roots, demonstrating that the total surface of root was the best parameter for the biomass prediction with high linear regression correlation (R² = 0.8312), which was sufficient for subsequently estimating the production of bioactive metabolites without content determination. This study provides an important approach for further grading of medicinal materials and breeding practices.